Claude review: kunit: Add support for suppressing warning backtraces

[Claude Code Review Bot <claude-review@example.com>]

Overall Series Review

Subject: kunit: Add support for suppressing warning backtraces
Author: Albert Esteve <aesteve@redhat.com>
Patches: 9
Reviewed: 2026-05-16T09:18:00.257637

---

This is a well-evolved v13 series (originally RFC by Alessandro Carminati, carried forward by Guenter Roeck, now stewarded by Albert Esteve) adding KUnit warning backtrace suppression. The design is solid: task-scoped suppression integrated into the existing KUnit hooks infrastructure, with zero overhead when no tests are running (via the `kunit_running` static branch). The series is cleanly split into infrastructure, tests, a real-world user, and documentation.

**Strengths:**
- Clean integration into existing KUnit hooks — no new config options, no arch-specific code
- Three suppression points covering both `__WARN_FLAGS` and non-`__WARN_FLAGS` architectures
- The `count` parameter to `kunit_is_suppressed_warning()` elegantly solves the double-counting problem
- The `in_task()` guard prevents false suppression from hardirq context
- Good test coverage including cross-kthread isolation
- Scoped API using `__cleanup` attribute is ergonomic

**Concerns (minor):**
- The `synchronize_rcu()` in cleanup path could add measurable latency in test suites with many suppression scopes, though it's correct
- The nesting prohibition (only one active suppression per task) may be limiting for complex test helpers, but is reasonable for v1

Overall this looks ready to merge with the reviews it already has (David Gow, Kees Cook). A few minor observations follow.

---

---
Generated by Claude Code Patch Reviewer